Wednesday, 23 June 2010
11: Mackenzie: Beliefs and Desires in Data
Contributions by Eleni Ikoniadou and Satinder Gill also showed the ways in which experimental art and performance could disrupt and reveal assemblages and hierarchies of the digital and the embodied in ways not open to conventional theoretical critique. It was, however, a paper by Adrian Mackenzie that I found most inspiring. Mackenzie considered the potential of a new materialist analysis of data networks. Taking as his starting point Gabriel Tarde’s formulation that all associations are intersections between different desires and beliefs, Mackenzie asked what could be learnt by considering vectors of belief and desire in data. In emphasizing the ‘in’, Mackenzie wants to go beyond a discussion of trust of data or beliefs about data onto analysis of the ways that beliefs and desires are invested in and in turn vitalise data. These beliefs and desires are not just embodied in users but also systems (e.g. Data Centres that seek to measure and respond to our desires).
Topographical models of distributed networks are, according, Mackenzie inadequate to understand the flux and dynamism of data movements and the resistances to those movements. Metaphors of ‘drag’, ‘lift’ ‘vortex’ seem more appropriate. Mackenzie examines the discourse of ‘data deluge’ that suggests that key contemporary problems relate to the containment and management of too much data – how to ‘mine’ for particular information, how to manage technical capacities, how to find patterns and order in the deluge? Mackenzie considered ‘the next big thing’ in data analysis and data mining the statistical programming language R. He discussed the energy and hope invested/expressed in R (e.g. the use of visualizations to bring data to life) as expressions of the plurality of desires and beliefs in data.
Mackenzie closed his paper by suggesting that we should stop thinking that data ever exists in a pure or raw form – it is given only in belief and desire.
Reflecting on my own work, this approach gave me another way of thinking about the part that data plays in contemporary politics and policy – the intersections of beliefs and desires in forensic DNA databases for example. But I was also left with some questions:
• Is there value in unpacking ‘beliefs and desires’ e.g. exploring ‘fears’ and 'hopes' in data might be useful?
• Perhaps also we should also consider‘beliefs and desires’ in different forms of data organization, management and representation e.g. categorization, abstraction, quantification, aggregation and visualization.
• There is also an interesting contrast (or perhaps I cannot see the join) between Mackenzie’s use of Tarde and Bruno Latour’s use of Tarde to champion possibilities of a non-reductive (and easily assessable) statistics of association which I have discussed previously.